Noninsulin-based antihyperglycemic medications in patients with diabetes and COVID-19: A systematic review and meta-analysis

1 INTRODUCTION

Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, is a worldwide pandemic that has caused more than 586 million cases and 643 million deaths (as of August 2022).¹ Clinically, COVID-19 can manifest as a wide range of symptoms, from asymptomatic to severe. Diabetes is associated with an increased risk of developing more severe COVID-19 and a greater risk of hospitalization, intensive care unit (ICU) admission, and mortality.^{2, 3} Furthermore, recommended treatment with corticosteroids increases the risk of uncontrolled glycemia.² The decision to continue using antihyperglycemic medications (AGMs) in patients with diabetes is challenging. This meta-analysis aims to study the impact of different groups of AGMs on patients with COVID-19. A comparison was made between those receiving the AGMs and those not receiving them.

2 METHODS

We searched Medline/PubMed, Embase, Web of Science, and Cochrane on January 24, 2022. We used the following keywords (COVID-19 OR Coronavirus disease 2019 [MESH]) AND (Diabetes mellitus) AND (Oral antihyperglycemic agent OR sulfonylurea derivative OR Dipeptidyl peptidase-4 inhibitor Or Sodium-glucose cotransporter 2 Or Empagliflozin Or dapagliflozin OR Canagliflozin OR Metformin OR Glucagon-like peptide-1 receptor agonist pioglitazone OR hypoglycemic agents).

We included in this analysis studies that included patients with type 2 diabetes (T2D) who were infected with COVID-19. The present meta-analysis examines the effect of different AGMs on mortality, hospitalization, and admission to the ICU and/or mechanical ventilation. Each category of AGM is compared with the other groups. We included studies that reported one or more of the following outcomes: mortality, hospitalization, ICU admission, or mechanical ventilation. We excluded systematic reviews, non-English articles, case reports, case series, literature reviews, posters, abstracts, pediatric and pregnancy cases, and patients receiving insulin for outpatient treatment. The term “mortality” was defined as death occurring within 30 days of confirmed infection with COVID-19. The identified articles were uploaded to the Covidence website for removal of duplicate files and screening. Coauthors independently screened, reviewed, and extracted the data from the full text. We extracted the data into a spreadsheet in Microsoft Excel. An evaluation of the risk of bias was carried out by two independent coauthors using the GRADE assessment for randomized controlled trial (RCT) (Figure S25) and Newcastle-Ottawa quality assessment for retrospective cohort studies (Table S1). The study did not involve human or animal subjects.

3 RESULTS

A total of 1365 articles were found after removing 144 duplicates. A total of 1221 articles were screened based on the title and abstract. A total of 1086 articles were excluded at this stage. We screened 135 articles based on their full text, and 108 articles were excluded. This study included 26 articles, shown in Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Figure 1 and Table 1.

4 DISCUSSION

We conducted a meta-analysis based on pooled data from 26 primary studies.^4-29 A higher risk of severe COVID-19 and mortality is associated with comorbidities such as cardiometabolic disease and diabetes.^{30, 31} Several AGMs have been studied, including etformin, DPP-4i, sulfonylureas, glinide, SGLT-2i, and GLP-1RA. Statistically significant findings include the following: metformin and GLP-1RA use were associated with statistically significant reductions in mortality risk, DPP-4i use was associated with a statistically significant increase in the likelihood of hospitalization and admission to ICU, and SGLT-2i use was associated with a statistically significant reduction in hospitalization risk. These findings have important implications for clinical practice. However, more robust studies, particularly RCTs, are needed.

In the metformin group, pooled analysis of 15 included studies revealed its use to be associated with decreased mortality compared to nonuse of metformin.^{5-7, 10, 12, 15-17, 19, 22-25, 27, 28} Our results agree with the findings of Scheen,³² who noted that metformin has a complex mechanism of action, some of which leads to anti-inflammatory activity that may help reduce the risk of severe COVID-19 beyond the effects of glucose control.^{10, 33} Nuclear factor kappa-light-chain-enhancer of activated B cells nuclear factor-kB (NF-κB) and mammalian target of rapamycin are two other pathways that may be affected by metformin to reduce the release of inflammatory markers.³² Further, metformin affects SARS-CoV2 replication and virus entry into the cell by phosphorylating the angiotensin-converting enzyme 2 (ACE2) receptor via adenosine monophosphate-activated protein kinase.^{34, 35} Finally, metformin has a role in other viral infections, such as stimulating autophagy in neuroblastoma cells and reducing herpes simplex virus type 1 (HSV-1) propagation.³⁶ In patients with dengue infection, metformin regulates immune-metabolic activities.³⁷

A mechanism for the observed increase in the risks of both hospitalization and ICU admission and/or mechanical ventilation in patients using DPP-4i compared to those not using DPP-4i is not entirely apparent. There may be the proinflammatory activity of DPP-4i, which might contribute to the hyperinflammation associated with COVID-19,³⁸ although others have suggested that DPP-4i exhibits direct anti-inflammatory, immunomodulatory, and antifibrotic effects.^39-41 As a receptor, DPP-4 may play a role in SARS-CoV-2 entry into the host.⁴²

Although the spike protein of the SARS-CoV-2 virus does not interact with human membrane-bound DPP-4 (CD26),^{43, 44} patients with T2D are thought to have dysregulated DPP-4 levels, which might have a negative vascular impact, resulting in an increased risk of COVID-19.⁴⁵ DPP-4 is present in visceral fat, contributing to insulin resistance and adipocyte inflammation, so that DPP-4i might correct some of the factors linking obesity with poor outcomes.⁴⁶

Many types of immune cells express DPP-4, including CD4(+) and CD8(+) T cells, B cells, NK cells, dendritic cells, and macrophages.^{44, 47} By activating the NF-kB signaling pathway, DPP-4 promotes the activation and proliferation of T cells.^{48, 49} There is evidence that DPP-4's effects are triggered by interactions between antigen-presenting cells and markers such as CD45, caveolin-1, mannose-6 phosphate receptors, or adenosine deaminase (ADA). Owing to its interaction with ADA, DPP-4 facilitates immune cell migration and diapedesis.⁴⁵ As a result of reducing cytokine storms, DPP-4i plays a role in preventing acute respiratory distress syndrome.^{48, 50, 51} In addition, DPP-4i has also been speculated to have antifibrotic effects because DPP-4 is known to stimulate the production of cytokines and chemokines by fibroblasts and the proliferation of smooth muscle cells. Therefore, DPP-4i may prevent lung fibrosis progression and reduce mechanical complications associated with COVID-19.⁴⁴

Based on a pooled analysis of five studies,^{7, 19, 25, 27, 28} there were no statistically significant differences in mortality between patients treated with sulphonylureas/glinides and those treated with non-sulphonylureas. Other studies suggest that sulfonylureas used prior to admission were associated with a borderline increased risk of adverse outcomes during hospitalization.⁵²

Some studies recommended that SGLT-2i should be temporarily discontinued in hospitalized patients because of the possibility of euglycemic ketoacidosis and dehydration during COVID-19 infection.⁵³ Two studies included in the pooled data analysis of hospitalization in the SGLT-2i group vs the non-SGLT-2i group revealed an association of decreased incidence of hospitalization in the SGLT-2i group compared with the non-SGLT-2 group.^{11, 13} There was no statistically significant effect of SGLT-2i use as regards ICU admission/mechanical ventilation and mortality. This finding is reinforced by an RCT comparing dapagliflozin to placebo among persons hospitalized with COVID-19 finding that dapagliflozin was not associated with improvement or adverse events.¹⁴ In patients with COVID-19, SGLT-2i may improve cardiovascular risk factors, including blood pressure, ambient glucose levels, weight, and cardiac function, along with anti-inflammatory effects, and it is possible that SGLT-2i might affect viral entry and infection.^{52, 54}

There are multiple and pleiotropic effects of SGLT-2i, including an improved endothelial function that may contribute to the reduction of thromboembolic complications as well as anti-inflammatory effects reducing inflammation markers like interleukin 6 (IL-6), ferritin, or C-reactive protein and reducing the intensity of the cytokine storm.⁵⁵

Our analysis showed that GLP-1RA was associated with a statistically significant reduction in mortality rate, ICU admission/mechanical ventilation, and hospitalization.^{11, 13, 21, 24, 25, 27, 28} Mechanism(s) for these beneficial effects are unclear but may be explained by or more of the following. It has been shown that GLP-1RA might reduce viral entry and infection in animal models.⁵² The effects of GLP-1 RA on chronic inflammatory diseases, including nonalcoholic fatty liver disease and atherosclerosis, are mediated by reduced activity of inflammatory pathways, which may affect COVID-19.⁵⁶ In a study on mice, GLP-1RA was shown to have an anti-inflammatory effect by reducing cytokine production and mucus secretion and preserving respiratory function.⁵⁷ GLP-1RA has antiobesity effects,⁵⁸ which may ameliorate low-grade chronic inflammation. Finally, it may improve obesity-associated decreased vitamin D bioavailability and gut microbiome dysbiosis.⁵⁹

GLP-1RA may also have antithrombotic and mitochondrial protective effects.⁶⁰ GLP-1 and ACE2 interact in a manner that has been the subject of considerable debate.^{60, 61} However, ACE2 upregulation induced by GLP-1RA may, paradoxically, ameliorate lung injury during COVID-19 despite enabling virus entry into host target cells.^59-61 Through binding to GLP-1R, GLP-1RA inhibits protein kinase C and NF-kB activation, decreasing the expression of NOD-like receptors (NLRs) family pyrin domain containing 3 (NLRP3), IL-1β, tumor necrosis factor-α, IL-6, vascular cell adhesion molecule 1, interferon-γ, and monocyte chemoattractant protein-1.⁵⁹

A strength of our analysis is the inclusion of different AGMs used during COVID-19. Second, our pooled analysis compares each group with all other groups. Third, our analysis included data of patients from a wide range of countries. This study has several limitations. The results of the analysis indicated moderate to high heterogeneity. The baseline level of glycosylated hemoglobin before infection with COVID-19 was not reported in the included studies. Limited information is available regarding other comorbidities, such as chronic obstructive pulmonary disease, obesity, asthma, cardiovascular diseases, and renal function. Most of the included studies involved patients during the early stages of the COVID-19 pandemic when anticoagulation and steroid treatment were not widely used. The study protocol was not registered in the International Prospective Register of Systematic Reviews (PROSPERO).

COVID-19 infection can be prevented through infection control measures, vaccination, hand hygiene, mask wear, and maintaining social distancing. Controlling glycemic levels will help reduce the severity of COVID-19 infection. The analysis suggests that therapy such as GLP1-RA metformin and SGLT-2i should be considered in populations with a high prevalence of COVID-19.

5 CONCLUSIONS

Our analysis showed the use of metformin or GLP-1RA to be associated with decreased risk of mortality. There was an increased risk of ICU admission/mechanical ventilation associated with using DPP-4i and a decrease associated with using GLP-1RA. There was an increased risk of hospitalization associated with DPP-4i and a decrease with SGLT-2i and GLP-1RA. Because heterogeneity was moderate to high among the included studies, we cannot recommend discontinuing any group of AGMs. More prospective studies on AGMs with COVID-19 are needed.

AUTHOR CONTRIBUTIONS

Conceptualization: Anoop Misra, Zachary Bloomgarden. Screening and data extraction: Mahmoud Nassar, Hazem Abosheaishaa; data check: Awadhesh Kumar Singh; reviewing: Awadhesh Kumar Singh, Anoop Misra, Zachary Bloomgarden; statistical analysis: Mahmoud Nassar; writing the discussion: Mahmoud Nassar. Reviewing: Anoop Misra, Zachary Bloomgarden, and Awadhesh Kumar Singh.

FUNDING INFORMATION

No funding agent or sponsor for this article.

CONFLICT OF INTEREST

There is no conflict of interest to declare.

ETHICAL STATEMENT

No human or animal studies were involved in this study.

PATIENT AND PUBLIC INVOLVEMENT SUBSECTION

Patients and the public were not involved in any way in this study.